[SPARK-43940][SQL] Test the error class: BATCH_METADATA_NOT_FOUND

[deepujain]

Summary

Adds coverage for the missing batch metadata error path.

Change

HDFSMetadataLog.applyFnToBatchByStream now has a focused test that checks BATCH_METADATA_NOT_FOUND, the missing batch file parameter, and SQLSTATE 42K03.

Tests

JAVA_HOME=/opt/homebrew/opt/openjdk@17 PATH="/opt/homebrew/opt/openjdk@17/bin:$PATH" ./build/sbt "sql/testOnly org.apache.spark.sql.errors.QueryExecutionErrorsSuite -- -z SPARK-43940"

Fixes SPARK-43940

JIRA assignee for credit: deepujain

[deepujain]

Fixed the commit author metadata and force-pushed the same test-only tree. Local check already passed: JAVA_HOME=/opt/homebrew/opt/openjdk@17 PATH="/opt/homebrew/opt/openjdk@17/bin:$PATH" ./build/sbt "sql/testOnly org.apache.spark.sql.errors.QueryExecutionErrorsSuite -- -z SPARK-43940". CI is running again on the updated head.

[deepujain]

Rebased on current master. I reran JAVA_HOME=/opt/homebrew/opt/openjdk@17/libexec/openjdk.jdk/Contents/Home PATH="/opt/homebrew/opt/openjdk@17/libexec/openjdk.jdk/Contents/Home/bin:$PATH" SBT_OPTS="-Dspark.driver.bindAddress=127.0.0.1 -Dspark.driver.host=127.0.0.1" ./build/sbt "sql/testOnly org.apache.spark.sql.errors.QueryExecutionErrorsSuite -- -z SPARK-43940" locally; it passed. CI has restarted on the new head.

[uros-b]

Thank you @deepujain!